3D Freespace Detection

Konzulens:
Szemenyei Márton
Kölső konzuelns:
Continental
Tárgy:
Önálló laboratórium 1 - Irányítórendszerek főspecializáció, MSc Vill.
Önálló laboratórium 2 - Irányítórendszerek főspecializáció, MSc Vill.
Önálló laboratórium 1 - Vizuális informatika főspecializáció, MSc Info.
Önálló laboratórium 2 - Vizuális informatika főspecializáció, MSc Info.
Hallgatói létszám:
1
Folytatás:
Szakdolgozat / Diplomaterv
Leírás:
Description
One important environment sensing aspect for automated vehicles is the so called Freespace, i.e. the
clear driveable area. It defines the area an automated driving function can use for path planning, or
for evasion maneuvers in case of emergencies.
State-of-the-art in camera-based freespace detection is to use semantic segmentation (pixel labeling)
to identify the road area in the 2D image, and its boundary. 3D sensors, most commonly laser
scanners are used to detect elevated obstacles such as barriers or overhanging load.
Goal of this work is to investigate in doing both with a single monocular camera. Prior work exists at
Conti to estimate depth from monocular images, and for semantic segmentation, upon which this
work can be based.
 
Requirements
  • BSc in Computer Science or similar
  • Hands-on experiences with deep learning tools like Keras, Caffe, or similar.
  • Experience in machine learning and recurrent neural networks
  • Good mathematical and analytical background
  • Strong programming skills in Python
 
Work can be executed as a Master Thesis of project work over at least 6 months.